Omit duplicate values in one column and show the rest - select

Question

I am starting with SQL, so my question will be easy for you.
I have one table with many columns. Most important are [Test number] and [Start time]. [Test number] contains many duplicated rows and [Start time] contains always unique records. And now I want to show in results all table columns, where only the [Test number] with maximum [Start time] will be used.
I was able to create one result regarding the maximum [Start time]
SELECT MAX([Start time])
FROM [Testing_ADB_Overview].[dbo].[HTT_DB]
GROUP BY [Test number]
ORDER BY MAX([Start time])
but I do not know how can I apply it to see in results all the other columns. Can you please help me?
MF

Answer 1

Try to get the [Test number] for the last [Stat time]. You can do it in two ways:
Store in a variable (#MaxStatTime) the result of the max from your select. Make a query of the [Test number] with the following where condition [Start time] = #MaxStartTime and store it in another variable (#LastTestNumber). And just retrieve the rows from your table corresponding to #LastTestNumber.
Using the ROW_NUMBER() function, you can order descending your rows in the table by the [Stat time] column. Then by filtering the table for row_number() = 1, you can obtain the [Test number], and again retrieve the rows from your table.
I'm not that fond about the 1st solution, so I will skip it's implementation, but here is an implementation of the 2nd solution using common table expression to make it more readable:
IF OBJECT_ID('tempdb.dbo.#HTT_DB') IS NOT NULL DROP TABLE #HTT_DB
CREATE TABLE #HTT_DB
( [Test number] NVARCHAR(100)
, [Start time] NVARCHAR(100)
)
INSERT INTO #HTT_DB( [Test number], [Start time] )
VALUES
('AA' , '210525_090000')
,('BB', '210525_080000')
,('CC', '210525_070000')
,('BB', '210525_060000')
,('AA', '210525_050000')
;WITH OrderedTestNumber AS (SELECT [Test number], [Start time], ROW_NUMBER() OVER ( PARTITION BY [Test number] ORDER BY [Start time] DESC) RowNo FROM #HTT_DB)
SELECT htt.*
FROM #HTT_DB htt INNER JOIN OrderedTestNumber otm ON otm.[Test number] = htt.[Test number] AND otm.[Start time] = htt.[Start time]
WHERE otm.RowNo = 1
Output:
Test number
Start time
AA
210525_090000
BB
210525_080000
CC
210525_070000
I would strongly advice adding a primary key column to ta table [HTT_DB], let's say PK_Column for the sake of simplicity. The query would bi simplifed as:
;WITH OrderedTestNumber AS (SELECT PK_Column, ROW_NUMBER() OVER ( PARTITION BY [Test number] ORDER BY [Start time] DESC) RowNo FROM #HTT_DB)
SELECT htt.*
FROM #HTT_DB htt INNER JOIN OrderedTestNumber otm ON otm.PK_Column = htt.PK_Column
WHERE otm.RowNo = 1